Results Summary: random ILR Augmentation Methods - Pseudo-count \(0.5\)

1 Logistic regression with \(\text{L}_1\) penalty

1.1 Interactive Summary Table

The table below shows the mean of the selected performance metric for the chosen augmentation factor.

Performance metric

Augmentation factor

1.2 Boxplots of performance metrics

One tabset with a box plot corresponds to a single augmentation factor in the section below.

1.2.1 ROC AUC

1.2.1.1 Density default

k = 2

k = 3

k = 4

k = 5

1.2.1.2 Density = 0.1

k = 2

k = 3

k = 4

k = 5

1.2.1.3 Density = 0.5

k = 2

k = 3

k = 4

k = 5

1.2.2 Misclassification Rate

1.2.2.1 Density default

k = 2

k = 3

k = 4

k = 5

1.2.2.2 Density = 0.1

k = 2

k = 3

k = 4

k = 5

1.2.2.3 Density = 0.5

k = 2

k = 3

k = 4

k = 5

1.3 Impact of the augmentation factor on the predictive perfromance

The plots below show the impact of the augmentation factor on the selected performance metric. Each tabset corresponds to a given density of the skew-symmetric matrix used for augmentation.

1.3.1 ROC AUC

Default

0.1

0.5

1.3.2 Misclassification Rate

Default

0.1

0.5

2 Random Forest

2.1 Interactive Summary Table

The table below shows the mean of the selected performance metric for the chosen augmentation factor.

Performance metric

Augmentation factor

2.2 Boxplots of performance metrics

One tabset with a box plot corresponds to a single augmentation factor in the section below.

2.2.1 ROC AUC

2.2.1.1 Density default

k = 2

k = 3

k = 4

k = 5

2.2.1.2 Density = 0.1

k = 2

k = 3

k = 4

k = 5

2.2.1.3 Density = 0.5

k = 2

k = 3

k = 4

k = 5

2.2.2 Misclassification Rate

2.2.2.1 Density default

k = 2

k = 3

k = 4

k = 5

2.2.2.2 Density = 0.1

k = 2

k = 3

k = 4

k = 5

2.2.2.3 Density = 0.5

k = 2

k = 3

k = 4

k = 5

2.3 Impact of the augmentation factor on the predictive perfromance

The plots below show the impact of the augmentation factor on the selected performance metric. Each tabset corresponds to a given density of the skew-symmetric matrix used for augmentation.

2.3.1 ROC AUC

Default

0.1

0.5

2.3.2 Misclassification Rate

Default

0.1

0.5

3 XGBoost

3.1 Interactive Summary Table

The table below shows the mean of the selected performance metric for the chosen augmentation factor.

Performance metric

Augmentation factor

3.2 Boxplots of performance metrics

One tabset with a box plot corresponds to a single augmentation factor in the section below.

3.2.1 ROC AUC

3.2.1.1 Density default

k = 2

k = 3

k = 4

k = 5

3.2.1.2 Density = 0.1

k = 2

k = 3

k = 4

k = 5

3.2.1.3 Density = 0.5

k = 2

k = 3

k = 4

k = 5

3.2.2 Misclassification Rate

3.2.2.1 Density default

k = 2

k = 3

k = 4

k = 5

3.2.2.2 Density = 0.1

k = 2

k = 3

k = 4

k = 5

3.2.2.3 Density = 0.5

k = 2

k = 3

k = 4

k = 5

3.3 Impact of the augmentation factor on the predictive perfromance

The plots below show the impact of the augmentation factor on the selected performance metric. Each tabset corresponds to a given density of the skew-symmetric matrix used for augmentation.

3.3.1 ROC AUC

Default

0.1

0.5

3.3.2 Misclassification Rate

Default

0.1

0.5